[brightcove] Improve embed detection (closes #25674)

[youtube-dl] / youtube_dl / extractor / brightcove.py

diff --git a/youtube_dl/extractor/brightcove.py b/youtube_dl/extractor/brightcove.py
index 5e0c4bc3eb584d4571e2e29b3034a8aeab3102fb..2aa9f4782e0dfdb2b78225c2d1fe83a8568effe3 100644 (file)
--- a/youtube_dl/extractor/brightcove.py
+++ b/youtube_dl/extractor/brightcove.py
@@ -5,32 +5,34 @@ import base64
 import re
 import struct
 
-from .common import InfoExtractor
 from .adobepass import AdobePassIE
+from .common import InfoExtractor
 from ..compat import (
     compat_etree_fromstring,
+    compat_HTTPError,
     compat_parse_qs,
     compat_urllib_parse_urlparse,
     compat_urlparse,
     compat_xml_parse_error,
-    compat_HTTPError,
 )
 from ..utils import (
-    ExtractorError,
+    clean_html,
     extract_attributes,
+    ExtractorError,
     find_xpath_attr,
     fix_xml_ampersands,
     float_or_none,
-    js_to_json,
     int_or_none,
+    js_to_json,
+    mimetype2ext,
     parse_iso8601,
     smuggle_url,
+    str_or_none,
     unescapeHTML,
     unsmuggle_url,
-    update_url_query,
-    clean_html,
-    mimetype2ext,
     UnsupportedError,
+    update_url_query,
+    url_or_none,
 )
 
 
@@ -424,7 +426,7 @@ class BrightcoveNewIE(AdobePassIE):
         # [2] looks like:
         for video, script_tag, account_id, player_id, embed in re.findall(
                 r'''(?isx)
-                    (<video\s+[^>]*\bdata-video-id\s*=\s*['"]?[^>]+>)
+                    (<video(?:-js)?\s+[^>]*\bdata-video-id\s*=\s*['"]?[^>]+>)
                     (?:.*?
                         (<script[^>]+
                             src=["\'](?:https?:)?//players\.brightcove\.net/
@@ -553,10 +555,16 @@ class BrightcoveNewIE(AdobePassIE):
 
         subtitles = {}
         for text_track in json_data.get('text_tracks', []):
-            if text_track.get('src'):
-                subtitles.setdefault(text_track.get('srclang'), []).append({
-                    'url': text_track['src'],
-                })
+            if text_track.get('kind') != 'captions':
+                continue
+            text_track_url = url_or_none(text_track.get('src'))
+            if not text_track_url:
+                continue
+            lang = (str_or_none(text_track.get('srclang'))
+                    or str_or_none(text_track.get('label')) or 'en').lower()
+            subtitles.setdefault(lang, []).append({
+                'url': text_track_url,
+            })
 
         is_live = False
         duration = float_or_none(json_data.get('duration'), 1000)
@@ -589,6 +597,7 @@ class BrightcoveNewIE(AdobePassIE):
         policy_key_id = '%s_%s' % (account_id, player_id)
         policy_key = self._downloader.cache.load('brightcove', policy_key_id)
         policy_key_extracted = False
+        store_pk = lambda x: self._downloader.cache.store('brightcove', policy_key_id, x)
 
         def extract_policy_key():
             webpage = self._download_webpage(
@@ -610,7 +619,7 @@ class BrightcoveNewIE(AdobePassIE):
                     r'policyKey\s*:\s*(["\'])(?P<pk>.+?)\1',
                     webpage, 'policy key', group='pk')
 
-            self._downloader.cache.store('brightcove', policy_key_id, policy_key)
+            store_pk(policy_key)
             return policy_key
 
         api_url = 'https://edge.api.brightcove.com/playback/v1/accounts/%s/%ss/%s' % (account_id, content_type, video_id)
@@ -638,6 +647,7 @@ class BrightcoveNewIE(AdobePassIE):
                         self.raise_geo_restricted(msg=message)
                     elif json_data.get('error_code') == 'INVALID_POLICY_KEY' and not policy_key_extracted:
                         policy_key = None
+                        store_pk(None)
                         continue
                     raise ExtractorError(message, expected=True)
                 raise